Chapter V · Limits of functions of a continuous variable. Continuous and discontinuous functions
Implicit and inverse functions
When an equation really defines as a continuous function of , and why a continuous strictly monotone function has a continuous inverse.
Why this matters
In Chapter II Hardy spoke freely of “the function defined by ”, and of , and their like, without asking whether such functions exist. An equation may have no solution for some , or several. In §§109–110 he settles the question near a point: if is continuous and strictly monotone in , the equation defines exactly one continuous function through the given point. The special case gives inverse functions, which we need for , , and above all for the logarithm as the inverse of the exponential (Chapter IX).
The idea
Hardy’s argument is a picture (his Fig. 33). Suppose and that, in a square round , is continuous and strictly increasing in for each fixed .
- Go straight up from to a point and straight down to . Since increases in , .
- By continuity these signs persist a little to the left and right: and for all near .
- On each such vertical line, is continuous and goes from negative to positive, so it vanishes somewhere (the intermediate value theorem), and being strictly increasing it vanishes only once. Call that point .
Try this. Drag the purple handle to move the vertical line across the square, and watch the slice on the right. Note where the signs at and stay opposite (the dashed band ). Then choose and do the same.
- f > 0
- f < 0
- f = 0
- square |x|, |y| ≤ k; dashed: |x| ≤ h
- the line x = x₁ (drag the handle) and y ↦ f(x₁, y)
What to notice
For the first two equations the slice is a strictly decreasing curve, so it crosses exactly once while and differ in sign; the crossings trace out the continuous curve . For the slice is a U-shape: two crossings when , none when . Monotonicity in is exactly what fails, and with it uniqueness and existence.
Precise statements
Hardy uses the word neighbourhood: a condition holds “in the neighbourhood of ” if it holds throughout some square centred at , and “in the neighbourhood of ” if it holds in some interval round .
Suppose that, in the neighbourhood of ,
- is a continuous function of and ;
- ;
- for each near , is a strictly increasing function of (or strictly decreasing).
Then, in the neighbourhood of , there is a unique function with values near (in the square) such that identically; it satisfies , and it is continuous in the neighbourhood of .
Hardy points out that the neighbourhood in the conclusion is generally smaller than the square in the hypotheses (in the widget, is usually less than ).
Taking , which is strictly monotone in exactly when is, gives a local inverse; Hardy then uses Heine–Borel to make it global, and finally gives a direct proof.
If is continuous and strictly increasing for , with and , then there is a unique function on with ; it is continuous and strictly increasing from to . The same holds with “decreasing” throughout.
Geometrically, the graph of is the graph of reflected in the line .
Try this. For , drag the purple point along the curve and watch its mirror image. Then choose and , and change the interval until the reflected curve becomes the graph of a function.
- the curve (y, F(y)), drawn with y across
- its reflection (F(y), y): the inverse, if it is a graph
- the line of reflection
- the level x = F(y) and its solutions
What to notice
Where is strictly monotone the reflected curve passes the vertical-line test: each has one , and the inverse is continuous. is not monotone on any interval round ; restrict to and you get , to and you get . is decreasing on , giving the unique solution of that vanishes with ; on a wider interval some levels are taken three times.
Proofs
Build by Hardy’s construction (existence and uniqueness on each vertical line), then get continuity by running the same construction in a smaller square: if the square has half-side , the solution is trapped within of .
Signs at the top and bottom
Take increasing in and a square where the hypotheses hold. Strict monotonicity in at gives .
The signs persist sideways
is continuous at and , and is continuous in (continuity in both variables implies it). By persistence of sign (§102) there is with and for .
Exactly one zero on each vertical
Fix with . The function is continuous on , negative at the bottom and positive at the top, so by §101 it vanishes at some ; being strictly increasing, it vanishes only once. Call that value . Then , , and no other in works.
Continuity at
Given with , repeat the first two steps with in place of : there is with for . For such the unique zero in lies in . So when : is continuous at .
Continuity at nearby points
For , the point satisfies the same hypotheses (in a smaller square inside the first), so the same argument shows is continuous at . Hardy reaches the same conclusion through the upper and lower limits , of at : both satisfy , so both equal .
For a level between and , the solution of should be the last at which is still at most : take to be the upper bound of . Then show , uniqueness, monotonicity, and continuity.
Define
Let and . contains and is bounded by , so it has an upper bound .
If , then by continuity on an interval , so would already be an upper bound of (no there or beyond has , since increases): contradiction. If , then by continuity on some , so contains points beyond : contradiction. (The end values give .)
Uniqueness and monotonicity
If then , since is strictly increasing. Write . If then : otherwise would give .
Continuity
is strictly increasing on and its values fill the whole interval (each in is ). By the theorem at the end of §101, a monotone function that takes every value between its end values is continuous.
Worked examples and exercises
Show that satisfies the hypotheses of the implicit function theorem at , with “decreasing” in condition 3.
Hint
Solution
is a polynomial, so it is continuous, and . For condition 3,
If , the bracket is at most , so has the sign opposite to : is strictly decreasing in . Hence exactly one continuous function satisfies the equation identically near and vanishes at .
For , check the hypotheses at and show that the function given by the theorem is . Why is the other root not the answer?
Hint
Solution
, and the second factor is negative when are small, so is strictly decreasing in near . As a quadratic in , has roots . At these are and . The theorem’s vanishes at and is continuous, so it is the root with the minus sign (with the positive square root); the other root is near , outside any small square round .
Discuss the equations and ( a positive integer) near , , in the light of §110.
Hint
Solution
is continuous and strictly increasing on the whole line (if then has the sign of , as is clear when have the same sign and trivial otherwise). So has a unique continuous inverse, , defined for all and vanishing at .
decreases for and increases for , so it is monotone on no interval round . Correspondingly has two solutions for and none for . Restricting to (where is strictly increasing) gives the inverse on .
As increases from to , is continuous and strictly increasing from to . Deduce the existence of a continuous, strictly increasing function on with .
Hint
Solution
Apply the theorem of §110 with on , , . It gives a unique function on with and , continuous and strictly increasing from to . This is , the value of the many-valued “arc sin” that lies in . Among all solutions of , it is the one that vanishes with . (That is strictly increasing on this interval follows from , both factors being positive.)
Check yourself
Near , which equation does the theorem of §109 show defines a unique continuous with ?
is strictly increasing in for every (a sum of two strictly increasing functions), continuous, and at the origin. and are not monotone in near : the first has two solutions or none, the second has the two solutions , both continuous. holds only at the origin.
If is continuous and strictly increasing on , its inverse is automatically continuous: no extra hypothesis is needed.
True: that is part of the theorem of §110. The inverse is strictly increasing and takes every value in , and a monotone function with no gaps in its values cannot jump (§101).
How many real solutions does have? (This is the level in the reflection widget with a wide interval.)
, with real roots , , . So on a wide interval the level is taken three times, and has no single inverse; only on an interval where it is monotone, such as , is the solution through unique.
Hardy’s language
There is nothing to show that the of the conclusion is the of the hypotheses, and indeed this is generally untrue.
Summary
- If is continuous near , , and is strictly monotone in , then defines exactly one function near with , and it is continuous.
- Proof: opposite signs at the top and bottom of a square persist sideways; on each vertical the intermediate value theorem gives a zero and strictness makes it unique; shrinking the square gives continuity.
- A continuous strictly monotone on has a continuous strictly monotone inverse on : its graph is the reflection of ’s graph in .
- Without monotonicity the conclusion fails: has two solutions or none near .